Case 4-2010 — A 53-Year-Old Man with Arthralgias, Oral Ulcers, Vision Loss, and Vocal-Cord Paralysis
 

John H. Stone, M.D., M.P.H., George N. Papaliodis, M.D., Mark R. Dunbar, M.D., and James R. Stone, M.D., Ph.D.

 

Presentation of Case

 

Dr. Arezou Khosroshahi (Rheumatology): A 53-year-old man was seen in the rheumatology clinic of this hospital because of arthralgias, oral ulcers, vision loss, and vocal-cord paralysis.

The patient had been well until 19 months before presentation, when he awoke with severe joint pain. On examination by his internist 1 month later, there was incomplete fist closure and decreased extension of the right elbow, without joint deformities or synovial swelling. Serum levels of electrolytes, bilirubin, calcium, thyrotropin, creatine kinase, C-reactive protein, and uric acid; the erythrocyte sedimentation rate; and tests of renal and liver function were normal. Testing for serum antinuclear cytoplasmic antibodies (ANCA) was negative. Ten days later, a rheumatologist detected a left-knee effusion. Synovial-fluid analysis revealed a white-cell count of 13,000 per high-power field (85% neutrophils), no crystals, and sterile cultures. Radiographs of the hands, wrists, knees, sacroiliac joints, and chest were normal. Assays for antinuclear antibodies (ANA), rheumatoid factor, and antibodies to cyclic citrullinated peptides were negative. Treatment with methylprednisolone, methotrexate, and folic acid were begun. The joint symptoms improved. During the following months, increased fatigue developed whenever methylprednisolone was tapered below 10 mg per day.

Fourteen months before this evaluation, treatment with etanercept was begun. Two months later, joint swelling, myalgias, low-grade fevers, and tinnitus developed. During the next 2 months, increasing muscle pain and tingling developed. Tests for ANCA were negative; other laboratory-test results are shown in Table 1. Etanercept was discontinued. Eight months before this evaluation, Bence Jones proteinuria was detected (Table 1). Serum levels of protein and complement were normal, and protein electrophoresis with immunofixation showed no M component. A radiographic skeletal survey and computed tomography (CT) of the chest, abdomen, and pelvis showed neither lesions (sclerotic or lytic) nor other evidence of a malignant condition. Two weeks later, the white-cell count was 2400 per cubic millimeter, with an absolute neutropenia and lymphopenia. Methotrexate was discontinued. Pathological examination of a bone marrow aspirate and biopsy specimen showed increased iron stores and 5% plasma cells in the marrow but was otherwise normal. Methylprednisolone was tapered to 4 mg per day.

Four months before this evaluation, daily low-grade fevers developed in the evenings, with temperatures to 38.1°C, associated with malaise, myalgias, arthralgias (most painful in the morning on awakening and after 4 p.m.), sore throat, large ulcers on the tongue accompanied by tongue swelling, and ophthalmic migraines (wavy lines in the patient's vision that were not associated with headaches). The patient awoke frequently at night because of discomfort in his shoulder, hip, elbow, or fingers. Eszopiclone was prescribed, without improvement. On two occasions, methylprednisolone (48 mg in a single dose) was administered for the ophthalmic migraines, with resolution of the ocular symptoms for 1 week on each occasion. Daily fevers persisted.

Three months before this evaluation, tinnitus worsened and hearing decreased. Auditory testing revealed high-frequency hearing loss in the right ear, which was attributed to methotrexate. Four days later, left-sided chest pain developed. The patient went to the emergency department of another hospital, where radiologic studies revealed a left pleural effusion with atelectasis and a thickened pericardium. Methylprednisolone (125 mg) was administered intravenously, followed by 40 mg orally, which was tapered to discontinuation over a period of 9 days. The chest pain resolved. Methotrexate was restarted. The following day, severe mouth ulcers and dysphagia developed. Treatment with acyclovir was begun, methylprednisolone was restarted, and symptoms improved within 24 hours.

Later that week, a neuro-ophthalmologist prescribed propranolol (20 mg twice daily) for ophthalmic migraine. The same day, the patient saw an otolaryngologist for severe throat pain and odynophagia associated with swallowing. On examination, there was an epiglottic ulcer; valacyclovir was administered, and methylprednisolone was tapered to 1 mg per day. Four days later, the patient saw rheumatologic and infectious-disease consultants because of persistent tongue ulcers and daily fevers. Cytomegalovirus cultures were negative.

The next evening, loss of vision occurred suddenly in the left eye, spreading in diagonal patches from the nasal to the temporal region, with total blindness in the left eye for approximately 30 minutes, then partial recovery in the inferior visual field. The patient was admitted to another hospital, where the results of funduscopic examination were interpreted as consistent with occlusion of the central retinal artery. Testing for ANCA was negative; other laboratory results are shown in Table 1. Magnetic resonance angiography of the brain and neck was normal. Methylprednisolone was stopped, and treatment with prednisone (60 mg per day) and a daily baby aspirin was begun. Transthoracic and transesophageal echocardiograms were normal. Examination of a biopsy of the right temporal artery was interpreted as negative for giant-cell arteritis by one pathologist but as "healed arteritis" by another. The patient's musculoskeletal symptoms and headaches improved, but the visual-field loss remained.

Nine days after the central retinal artery occlusion and 2 months before this evaluation, hoarseness developed abruptly. Otolaryngologic examination disclosed right vocal-cord paralysis. Magnetic resonance imaging of the head and neck revealed swelling of the right vocal cord. Three weeks later, the patient's dentist noted ulcers on the tongue up to 1 cm in diameter and 0.5 cm deep. Methotrexate and leucovorin were discontinued, and amoxicillin was begun. Six days later, oral fluconazole and parenteral triamcinolone (80 mg into the tongue) were administered. During the next 9 days, odynophagia and additional painful tongue lesions developed. The patient resumed treatment with acyclovir. Fourteen days before this evaluation, administration of acyclovir was stopped and famciclovir started.

Eight days before this evaluation, a cardiac stress test was performed but was stopped after 12 minutes because of sudden throbbing bitemporal headache. On examination, the temporal arteries were not tender. The headache resolved gradually within 30 minutes, without medication.

On evaluation in the rheumatology clinic of this hospital, the patient reported mild temporomandibular joint pain and recurrent episodic diplopia; the diplopia had been attributed to an acquired Brown syndrome (transient paralysis or tendinopathy of the superior oblique muscle). He had had ulcerative proctitis decades earlier. The patient was an ophthalmologist. He was married and had two children, who were well. There was no family history of cancer or autoimmune diseases. He was allergic to honeybee venom. Medications included prednisone (20 mg in the morning and 15 in the evening), diclofenac, propranolol, acetylsalicylic acid (81 mg), viscous lidocaine mouthwash, sucralfate suspension for mouth ulcers, folic acid, zinc, a multivitamin, and calcium with vitamin D.

On examination, the patient appeared well, but his voice was faint. The vital signs were normal. There was a well-healed scar from a temporal-artery biopsy on the right, and there were normal temporal-artery pulsations bilaterally. There were large, bilateral noninflamed ulcers of the tongue and small erosions on the top of the tongue posteriorly (Figure 1). The remainder of the examination was normal. Serum levels of electrolytes, calcium, total protein, albumin, globulin, total bilirubin, complement, and serum free light chains and tests of coagulation and renal and liver function were normal. Testing for antibodies to the Sm and U1-RNP antigens, β2-glycoprotein I, cyclic citrullinated peptide, rheumatoid factor, anticardiolipin antibodies, and lupus anticoagulant was negative. Serum immunoelectrophoresis was normal. Other results are shown in Table 1.

Figure 1. Tongue Ulcer.

A photograph of the patient's tongue at presentation, after he had been treated for months with variable doses of prednisone and had had some healing, shows a deep ulcer with raised edges. A similar lesion was present on the other side of the tongue.

 

 

 

A diagnostic procedure was performed.

Differential Diagnosis

Dr. John H. Stone: The patient in this case was Dr. Mark Dunbar. Dr. Dunbar, would you describe in detail your symptoms during the episode of blindness?

Dr. Mark R. Dunbar: The vision loss occurred at 7:07 p.m., while I was out at dinner. I suddenly noticed distortion in my vision. Alternate covering of the eyes revealed a painless, patchy, diamond-shaped loss of vision marching across my left visual field, nasally to temporally (see Video, available with the full text of this article at NEJM.org). The event lasted approximately 30 seconds. When it was over, I had no light perception in my left eye. I started intermittent digital pressure and paged the ophthalmology resident on call at the hospital where I work. The resident noted a 3+ Marcus Gunn pupil (decreased constriction of both pupils when a light is shone into the left eye), indicating a relative afferent pupillary defect, and confirmed the presence of a central retinal artery occlusion, which I had feared. We continued intermittent digital pressure, and at 7:35 p.m., blurry black-and-white vision returned in the lower half of my left visual field. Funduscopic examination at that time showed that the clot had moved from the central retinal artery to the inferior branch artery. By 8 p.m., color vision had returned to the lower visual field, though mild blurriness remained. The entire superior field remained black until the next day, when peripheral light perception returned.

Dr. J.H. Stone: Dr. Papaliodis, may we review the images of the patient's ocular fundus?

Dr. George N. Papaliodis: A photograph of the fundus approximately 40 hours after the onset of visual loss shows an occlusion of the inferior branch of the retinal artery in the left eye, with associated retinal edema (Figure 2A). There is no visualization of retinal emboli. The superior retina appears normal, which is consistent with the spontaneous recovery of the patient's inferior visual field. Fluorescein angiography performed at the same time (Figure 2B) shows a thrombus in the inferior branch of the retinal artery.

Figure 2. The Left Optic Fundus Approximately 40 Hours after the Onset of Vision Loss.

A color photograph of the fundus (Panel A) shows evidence of an occlusion of the inferior branch of the retinal artery, with a clot causing narrowing at both the first and second bifurcations (arrowheads) and retinal edema in the lower half of the retina (arrows). The superior retina appears normal, which is consistent with the spontaneous recovery of the patient's inferior visual field. An image from a fluorescein angiogram (Panel B) taken 5 minutes after the injection of dye shows a bright focus of mural thrombus (arrow). Perfusion was normal at the time of the angiography.

 

These findings, along with the initial notation of no light perception, suggest an occlusion of either the central retinal artery or the ophthalmic artery, with spontaneous restoration of perfusion to the superior retinal arcade, associated with recovery of his inferior visual field. Causes of central retinal artery occlusion include carotid atherosclerosis, cardiogenic embolism, hematologic diseases (e.g., sickle cell, hypercoagulable states, leukemias), and inflammatory diseases (e.g., vasculitides).

Dr. J.H. Stone: I am aware of the diagnosis in this case. This 53-year-old ophthalmologist had a central retinal artery occlusion. This life-altering event occurred on the background of inflammatory joint pains, tongue ulcers, vocal-cord paralysis, and fevers. The evaluation was marked by the finding of Bence Jones proteinuria. In addition, ANA assays changed from negative to positive and the titers increased with each measurement. A temporal-artery biopsy did not provide a definite diagnosis.

This patient's presentation has several potential "red herrings." His syndrome unfolded over a period of more than 18 months, during which time he was evaluated by 10 specialists. In addition, he had been taking immunosuppressive medications, which were likely to have altered the natural history of his disease. The essential painstaking review of the history was aided by meticulous notes that the patient had maintained to track his symptoms, test results, and medication doses.

Amyloidosis

The greatest urgency was to rule out amyloidosis. A urinalysis had revealed Bence Jones proteinuria, which results from the production of monoclonal immunoglobulin light chain by a clone of plasma cells.1,2 Amyloidosis results when a portion of the light-chain polypeptide forms a β-pleated sheet, which accumulates in tissues.3 Amyloidosis may cause macroglossia, which may be associated with tongue ulcers, and in view of its tropism for blood vessels, many other features of the patient's presentation could have been explained by AL amyloidosis. However, the patient did not have macroglossia, serum protein electrophoresis had not revealed an M component, a skeletal survey and CT scans of the chest and abdomen had shown no abnormalities, and a biopsy specimen of the bone marrow had revealed only 5% plasma cells.

Systemic Lupus Erythematosus and Rheumatoid Arthritis

Clinical features suggestive of systemic lupus erythematosus, including the increasing ANA titer, pleuropericarditis, and cytopenia, constituted the second red herring. The ANA test was negative at the time of the patient's first evaluation. Because he had inflammatory joint symptoms, the patient was eventually treated with etanercept, a tumor necrosis factor {alpha} (TNF-{alpha}) antagonist. TNF-{alpha} inhibitors can cause an array of autoimmune problems, including the syndrome of drug-induced lupus.4 The fact that this patient's ANA test was negative at his first evaluation but then became strongly positive and was associated with the presence of antihistone antibodies implicates drug-induced lupus as the cause of several features of his presentation. I believe that this same immunologic process led to the overproduction of immunoglobulin light chains, which in turn led to Bence Jones proteinuria. However, drug-induced lupus is usually characterized by skin rash, arthritis, serositis, and other relatively mild abnormalities. Therefore, it does not explain the entire clinical picture.

I considered the possibility that the patient had seronegative rheumatoid arthritis. His musculoskeletal symptoms, including symmetric arthralgias, as well as arthritis that involved distal and proximal joints, were consistent with rheumatoid arthritis. However, seronegative rheumatoid arthritis is unlikely to cause extensive extraarticular disease. Extraarticular rheumatoid arthritis (e.g., vasculitis) usually emerges only years after the onset of destructive arthritis and is unusual in a patient who lacks both rheumatoid factor and antibodies against cyclic citrullinated peptides.

Three forms of systemic vasculitis merit careful consideration. These include Behçet's syndrome, Wegener's granulomatosis, and giant-cell arteritis.

Behçet's Syndrome

Behçet's syndrome is characterized by oral and genital ulcerations.5 It can cause a host of ocular complications, including thrombotic events (e.g., central retinal artery or vein occlusion) or, more commonly, some form of uveitis (e.g., panuveitis or anterior, intermediate, or posterior uveitis).6 This patient had ocular abnormalities and oral ulcers. However, several points argue against Behçet's syndrome, the most important of which is the character of the oral lesions. Behçet's syndrome usually causes aphthous oral ulcers that are more superficial than this patient's large and deep lesions. Also, the aphthous ulcers of Behçet's syndrome can occur on the tongue but are equally likely to involve the buccal mucosa and soft palate, whereas in this patient, they occurred only on the tongue.7 Moreover, a central retinal artery occlusion is unlikely to occur in Behçet's syndrome in the absence of more classic ocular manifestations.8

Wegener's Granulomatosis

Wegener's granulomatosis can cause oral ulcers identical to those seen in this patient.9 It also can cause central retinal artery occlusions and vocal-cord lesions.10,11 However, the more common ocular lesions of Wegener's granulomatosis are scleritis and orbital pseudotumor, and the classic laryngotracheal lesion is subglottic stenosis.12,13 After 19 months of nearly continuous illness in this patient, the diagnosis of Wegener's granulomatosis would be improbable because of the absence of any evidence of disease in the kidneys, lungs, or sinuses; several negative tests for ANCA also argue against the diagnosis.

Giant-Cell Arteritis

The final major diagnosis to consider is giant-cell arteritis, a disease of large- and medium-sized arteries that tends to involve vessels of the head and neck. Giant-cell arteritis is seldom diagnosed in persons younger than 50 years of age,14 so this patient, at 53 years of age, is at the younger end of the disease spectrum. The diagnosis of giant-cell arteritis is tenable if the explanation for his prominent musculoskeletal discomfort is polymyalgia rheumatica, a musculoskeletal pain syndrome that classically involves the proximal areas of the shoulder and hip girdles and accompanies giant-cell arteritis in approximately 50% of cases.15 Several features of the patient's musculoskeletal pain suggest polymyalgia rheumatica, particularly the abrupt onset, dramatic response to glucocorticoids, and tendency to occur in the early morning. Distal symptoms were most prominent at presentation, but in some patients it is impossible to distinguish polymyalgia rheumatica from rheumatoid arthritis at the early stages of disease.16,17,18 Proximal as well as distal symptoms were present at the start of the patient's illness and remained a persistent feature of his pain. Patients with polymyalgia rheumatica seldom undergo arthrocentesis, but synovitis with joint effusions are well described17; the knee is among the most commonly involved joints. I believe that polymyalgia rheumatica explains the patient's musculoskeletal pain.

Anterior ischemic optic neuropathy, caused by lesions within the posterior ciliary artery, is the most common syndrome of vision loss in giant-cell arteritis. However, central retinal artery occlusion occurs in approximately 8% of patients with giant-cell arteritis who lose vision.19 Tongue ulcers in giant-cell arteritis were first described in 1962,20,21 but subsequently, many case reports have highlighted this feature.22 Hoarseness and vocal-cord paralysis are unusual but recognized manifestations of giant-cell arteritis,23 underscoring the precept that giant-cell arteritis must be suspected in patients in the appropriate age group who have any type of inflammatory lesion that occurs above the neck.

The patient also had sensorineural hearing loss, fevers, temporomandibular joint pain, and Brown syndrome, which fall within the spectrum of giant-cell arteritis. Sensorineural hearing loss is often overlooked among other more prominent symptoms of giant-cell arteritis, particularly in patients with baseline presbycusis.24 Fever and fever of unknown origin are common complications of giant-cell arteritis.25 The patient's temporomandibular joint pain may have represented jaw claudication, the feature of a medical history with the highest positive predictive value for a positive temporal-artery biopsy.14 Similarly, although acquired Brown syndrome is associated with diplopia caused by tendinopathies of the superior oblique tendon,26,27 giant-cell arteritis also causes diplopia by a variety of mechanisms: ischemia of the extraocular muscles, vasculitic neuropathy of the cranial nerves, and brain-stem ischemia.28

Finally, one must consider the patient's history of ophthalmic migraines. In contrast to classic migraines, ophthalmic migraines are not associated with headache but are transient, painless, and reversible visual phenomena. To my knowledge, such phenomena have not been described in giant-cell arteritis. However, giant-cell arteritis has been associated with hallucinations, namely the Charles Bonnet syndrome,29 in which visual hallucinations occur because of ischemia. It is easy to imagine that visual phenomena such as ophthalmic migraines could occur in giant-cell arteritis.

In summary, I believed that giant-cell arteritis was the most likely explanation for the patient's syndrome. I considered a variety of approaches to confirming the diagnosis, including a second temporal-artery biopsy and a CT angiogram of the great vessels. In the end, we decided to reexamine tissue that we already had, from the original temporal-artery biopsy.

Dr. Eric S. Rosenberg (Pathology): May we have the medical students' diagnosis?

A Harvard Medical Student: We considered a range of diagnostic possibilities, including Behçet's syndrome, amyloidosis, and giant-cell arteritis. We were convinced by the combination of oral ulcers, arthritis, and central retinal artery occlusion that the most likely diagnosis was Behçet's syndrome.

Dr. John H. Stone's Diagnoses

Giant-cell arteritis with central retinal artery occlusion.

Drug-induced lupus erythematosus.

Pathological Discussion

Dr. James R. Stone: The biopsy specimen of the tongue showed ulceration and necrosis that were consistent with ischemia. There was no evidence of invasive fungal infection or vasculitis, and a Congo red stain was negative for amyloid. The biopsy specimen of the temporal artery showed extensive intimal hyperplasia with fragmentation of the internal elastic lamina and focal scarring of the vessel wall (Figure 3A). Although it is not specific, such vascular scarring is consistent with healed vasculitis. Examination of the slides also revealed a sparse inflammatory infiltrate, primarily in the adventitia (Figure 3B). To further characterize the extent and nature of this infiltrate, immunohistochemical studies were performed. Multiple CD68+ macrophages were present within the artery wall around the internal elastic lamina (Figure 3C), and scattered CD3+ lymphocytes were present in the adventitia (Figure 3D). Giant cells were not observed. The inflammatory infiltrate was consistent with mild involvement by active giant-cell arteritis.

Figure 3. Temporal-Artery Biopsy.

A specimen from a cross section of the temporal artery at low magnification (Panel A, hematoxylin and eosin) shows extensive intimal hyperplasia and scarring of the wall (arrow). A high-power image (Panel B, hematoxylin and eosin) shows scattered lymphocytes in the adventitia. The macrophage marker CD68 (Panel C, immunoperoxidase) highlights multiple macrophages present at the internal elastic lamina (arrows). The lymphocytes are CD3+ T cells (Panel D, arrowheads; immunoperoxidase).

 

Giant-cell arteritis is a segmental disease with skip lesions, and thus histologic specimens may show a spectrum of severity from florid transmural involvement to more subtle involvement of the adventitia and internal elastic lamina. In cases of limited involvement, the adventitial inflammatory infiltrate may be predominantly lymphocytic, and that located at the internal elastic lamina may contain predominantly macrophages. To my knowledge, a multi-institutional consensus statement has not been devised by vascular pathologists to formally address the minimal diagnostic criteria for active giant-cell arteritis, and standards vary considerably among institutions. Although complete transmural inflammation is generally not required for a diagnosis of active giant-cell arteritis,30 some degree of involvement of the arterial media or intima, or both, is typically present. Giant cells are present in only about half of biopsy specimens of the temporal artery in patients with active giant-cell arteritis.31 Since subtle infiltration of the vessel wall by macrophages may not be readily appreciated on slides stained with hematoxylin and eosin, immunohistochemical staining for the macrophage marker CD68 may assist in the diagnosis of giant-cell arteritis, as it did in this case.32,33,34

Discussion of Management

Dr. J.H. Stone: The temporal-artery biopsy is the standard of diagnosis for giant-cell arteritis. However, there are many ways in which temporal-artery biopsies can yield false negative results (Table 2). In the diagnosis of giant-cell arteritis, the most critical component of the evaluation is the history. Without a carefully elicited medical history, the clinician will not suspect the diagnosis except in classic cases. In addition to the importance of the history, this case illustrates the importance of an experienced pathologist. Staining temporal-artery-biopsy specimens for CD68+ cells should be considered when the routine histopathological findings appear inconsistent with the clinical circumstances.

 

 


After the diagnosis of giant-cell arteritis was confirmed, we began a slow taper of the patient's prednisone over a period of 16 weeks, to a daily dose of 10 mg. We continued the taper at the rate of 1 mg per month. He is currently taking 8 mg per day, along with 325 mg of aspirin.44,45

Dr. Rosenberg: Dr. Dunbar, would you tell us how you are now?

Dr. Dunbar: I still practice ophthalmology but have decreased my schedule and stopped operating because of a persistent left paracentral scotoma. I continue my own speech therapy and have undergone two procedures on my paralyzed vocal cord. Tinnitus persists, but I am usually able to ignore it. Fatigue was a severe symptom initially but has largely resolved, and I have returned to swimming and jogging.

Dr. Rosenberg: Dr. Murali, would you tell us about additional laboratory testing?

Dr. Mandolathakur R. Murali (Immunology): Repeated analysis of the patient's urine 7 months after his initial evaluation at this hospital indicated that the small amount of Bence Jones protein detected initially was no longer present. The quantity of free kappa light chain in the urine diminished from 154.2 to 37.9 mg per liter, and the ANA titer from 1:2560 to 1:1280. These findings are consistent with an immune perturbation leading to an imbalance of the ratio of kappa to lambda light chains in the urine. The imbalance appears to be resolving, which is consistent with a resolving drug-induced lupus syndrome that presumably was caused by etanercept.

Anatomical Diagnosis

Temporal-artery findings consistent with mild involvement by active giant-cell arteritis.

 

 

This case was presented at the Medical Grand Rounds, May 21, 2009.